The discovery that lithium treatment at blastula stages can induce axis formation suggested that it might act by respecifying the cytoplasmic rearrangement-generated dorsoventral pattern, so that ventral cells behave like their dorsal counterparts. We have studied the effects of Li+ treatment on the spatial layout of the cell-group movements of gastrulation to see whether this is the case. We find that involution of the chordamesoderm and associated archenteron roof is retarded by Li+, an effect which does not suggest dorsal respecification. However, in both migration of the leading edge mesoderm and convergent extension of the marginal zone, ventral regions clearly do show dorsal-type movement. Because of this, and because of examples where disruption of involution and effects on axis differentiation do not correlate, we propose that failure of involution represents a distinct effect of Li+ involving disruption of mechanical relationships at the blastopore. Thus archenteron formation poorly reflects the dorsoventral pattern. Extension of sandwich explants of the ventral marginal zone is proposed as a reliable quantitative assay for alterations to the dorsoventral pattern.

In amphibians, formation of the body axis is thought to occur by means of a series of inductions (Nieuwkoop, 1973; Slack, 1983; Gurdon, 1987). A dorsal cue is created in the vegetal half of the fertilized egg by rearrangement of a preexisting cytoplasmic animal– vegetal gradient (Gerhart et al. 1984). During blastula stages, vegetal blastomeres induce a spatial pattern of morphogenetic and differentiative capacities in their animal cap neighbours, the marginal zone (Gimlich, 1986; Gimlich & Gerhart, 1984). During gastrulation, this spatial pattern of capabilities appears as a set of region-specific cell movements, which generate the body axis on the dorsal side of the embryo (Gerhart & Keller, 1986). In the process, mesodermal tissue of the head and notochord are brought under responsive ectoderm, in which they induce the head-to-tail neural axis (Suzuki et al. 1984; Spemann, 1938).

Perturbation of the dorsoventral pattern appears to affect the layout of gastrular cell-group movements and axial differentiative capacity in tandem. If the rearrangement of the egg cytoplasm is prevented by irradiation of the vegetal hemisphere with ultraviolet light (Scharf & Gerhart, 1983; Vincent et al. 1986), dorsal involution does not occur on schedule, but is delayed until the time appropriate for the ventral side (Malacinski et al. 1977; Scharf & Gerhart, 1980). The symmetrical, truncated shape of the resultant archenteron roof (C.M.R., unpublished data) reflects this ‘ventralization’. The mesoderm from ventralized embryos lacks the capacity to induce neural tissue from competent ectoderm (Malacinski et al. 1977) and differentiates exclusively into ventral, posterior derivatives (Grant & Wacaster, 1972). Thus, ventral differentiation and morphogenesis reflect the preexisting properties of the cytoplasm, in the absence of the cytoplasmic rearrangement.

Recently, Kao et al. (1986) reported that lithium treatment of embryos at blastula stages can both rescue anterior structures in ventralized embryos and cause reduction in posterior structures in normal ones. This finding suggested that, in contrast to previous workers’ characterization of Li+’s effect (e.g. Bäckström, 1954), Li+ might act to ‘dorsalize’ the pregastrular dorsoventral pattern, respecifying positional values so that ventral cells move and differentiate like their more dorsal counterparts. It is important to know whether this is truly the case.

Li+ undoubtedly confers dorsal anterior differentiative capacity. Ventralized (dorsal) mesoderm gains dorsal inductive capacity and normal ventral mesoderm is caused to differentiate as lateral or dorsal tissue (Kao et al. 1986). Effects on differentiation could be secondary, however, to a disruption of the normal spatial layout of gastrulation. The movements of gastrulation are the first manifestation of the dorsoventral pattern and so ought to provide a more direct test.

Previous reports on the effect of Li+ on gastrulation have not described the dorsoventral pattern of gastrular tissue movements. Li+-treated embryos generally exhibit a wide range of forms at stages corresponding to the end of gastrulation in controls, often involving extension of a long proboscis (Bäckström, 1954). Eggs treated with D2O, thought to dorsalize through its effects on cytoplasmic rearrangement, also exhibit these forms, suggesting that the proboscis might indeed reflect dorsaiization (Scharf et al. 1984).

However, we initially observed that Li+ retards dorsal involution and, in this respect, seems to ventralize. This result suggested that the relationship between morphogenesis and the dorsoventral pattern might be complex, and that Li+ might perhaps weaken the dorsal cues responsible for regulating gastrulation, thus respecifying dorsal and ventral tissue towards intermediate, lateral fates. It thus became imperative that the effect of Li+ on the spatial pattern of gastrulation be clarified.

We have used morphometric methods to analyse the pattern in gastrulation in Li+-treated animals, in order to determine whether the effect of Li+ on morphogenesis does represent dorsaiization and, if so, to establish a quantitative assay for the process. Our results indicate that Li+ treatment does indeed respecify ventral and lateral tissue toward more dorsal identities, but that it also causes disruption of mechanical relationships at the blastopore lip, so that involution is retarded. For this reason, the spatiotemporal layout of gastrulation, and thus the shape of the archenteron, poorly reflects the dorsalized positional values. On the other hand, the extension and convergence of the upper portion of the embryonic marginal zone reflects much more reliable respecification of ventral locations to perform dorsal region-specific movements. Sandwich explants from this region appear to provide a true quantitative assay for this dorsal respecification.

Eggs of Xenopus laevis were artificially fertilized, dejellied and reared in 20 % modified Steinberg’s solution (MSS) as previously described (Regen & Steinhardt, 1986).

Li+ treatment and scoring of axial anteriorization

Li+ treatments were performed at 32- or 64-cell stages. Embryos were transferred to a solution of 0·30M-LIC1 dissolved in 40% MSS (Kao et al. 1986) and, after an appropriate number of minutes, removed and washed several times in 20 % MSS. Embryos were usually reared at 15°C before and after the Li+ treatment. Axis-effect scores were determined by allowing embryos to develop to stage 40 (Nieuwkoop & Faber, 1967) and beyond, and then scoring for the degree of axial anteriorization (see Results).

Morphometric studies

It was necessary to mark the site of initial bottle cell pigment condensation to indicate the dorsal meridian. This was done by means of chips of Nile blue dye, in a modification of the method of Kirschner & Hara (1980). Nile blue sulphate (Sigma) was dissolved in distilled water, precipitated by dropwise addition of 1 M-Na2CO3, filtered and washed with more Na2CO3. Upon drying, the residue breaks into small, thin chips, which are then used for vital staining.

Embryos were transferred into hemispherical wells formed in 1·5 % agarose, bathed in 7 % Ficoll 400 (Sigma) dissolved in 40 % MSS, to dehydrate the perivitelline space (Kirschner & Hara, 1980). After a few minutes in this solution, the embryos can be stably positioned with their vegetal hemispheres upward. When pigment condensation first became visible, each embryo was rotated so that the dorsal meridian marked by this condensation was upward. After the majority of embryos were rotated, they were marked en masse, with small chips of dye. After 15 min, the chips were knocked off, leaving a dark spot to mark the dorsal meridian. After staining, embryos were transferred back to 20% MSS.

When they reached appropriate stages of development, embryos were fixed in 2 % glutaraldehyde/100mm-sodium cacodylate, pH 7·4. After fixation, they were transferred to glutaraldehyde-free cacodylate buffer, fractured with a blunt knife and photographed with a dissecting microscope. Some specimens were also dehydrated through an ethanol series, critical-point dried, mounted and sputter-coated with gold for scanning electron microscopy.

Morphometric parameters were determined from the light micrographs, by tracing the appropriate photographic feature with the digitizing arm of a Numonics graphics calculator interfaced to a computer. Typically, the mean of ten measurements was used to characterize a group of embryos.

Archenteron length was measured at stage 12, along the archenteron wall, from the highest extent of involution to the blastopore (inset, Fig. 6). Embryos were fractured midsagittally through the axis, and the path length of dorsal and ventral involution of the superficial endoderm determined.

Leading edge displacement was measured along the face of the yolk mass, from the blastopore to its upper vertex (inset, Fig. 5). Embryos were fixed for this measurement at stage , at which time the pigment line has nearly reached the ventral meridian, so that the marginal zone boundary is evident.

Gastrula-stage control embryos were staged according to blastopore diameter and archenteron length, using the data of Nakatsuji (1975).

The location and shape of the involuting archenteron roof were determined by fracturing embryos equatorially, and then pushing a blunt glass needle through the blastopore to dislodge the yolk mass. The pigmented archenteron roof could then be distinguished against the unpigmented blastocoel inner wall.

Histology

Embryos were soaked overnight in Smith fixative (Humason, 1979) then for an additional day in 01M sodiumcacodylate buffer. They were dehydrated through an ethanol series, soaked in Histosol (National Diagnostics, Somerville, NJ, USA) for 4h, and then embedded in paraffin. 8/im serial sections were stained with haematoxylin/eosin B/phloxine B and photographed using epifiuorescence with a Nikon Diaphot-TMD microscope. The excitation wavelength was 480 nm.

Sandwich explants

Sandwich explants of the marginal zone have been used by Keller and coworkers to analyse the morphogenetic movements of gastrulation (Keller et al. 1985). Explants were reared in 40 % MSS at room temperature and photographed every few hours, during the 8h period during which they extended. The first set of photographs were taken at a time before perceptible extension had occurred, defined to be t = 0. The time course of extension was determined by subtracting explant lengths in later photos from those in the original ones.

Explants were made as pictured in Fig. 1. When feasible, paired dorsal and ventral sandwiches were fashioned from the same two embryos. A working surface of Permaplast modelling compound was used. Embryos were mechanically decapsulated, and then divided into dorsal and ventral halves with an eyebrow-hair knife, in full-strength MSS, typically at 17°C. The remaining steps, which take about 10min to perform, are as follows:

Fig. 1.

Creation of sandwich explants of the marginal zone. See text for explanation. The lower, lighter region in the third step corresponds to the involuting marginal zone, while the upper darker region is composed of the noninvoluting marginal zone and part of the animal cap above it.

Fig. 1.

Creation of sandwich explants of the marginal zone. See text for explanation. The lower, lighter region in the third step corresponds to the involuting marginal zone, while the upper darker region is composed of the noninvoluting marginal zone and part of the animal cap above it.

  1. The top portion of the half-embryo’s blastocoel roof is sliced off.

  2. The yolk mass is removed from the blastocoel wall, by tugging at it with a baby-hair loop, while a hair knife probes the interface, to dislodge any remaining adherent cell protrusions. The marginal zone separates from the yolk mass near the blastopore lip.

  3. The marginal zone explant is cleaned up, by cutting the edges square and brushing off any head mesoderm cells protruding atop the smooth explant inner surface.

  4. Steps 1–3 are repeated for the second embryo half,

  5. A sandwich is formed. The first explant half, which may have contracted and rounded somewhat by this time, is inverted atop the second and pressed down to flatten it. Mounds are rubbed up in the substrate on two sides of the sandwich and a piece of coverslip pressed down on these, to squeeze the sandwich halves firmly together.

The sandwich is allowed to heal, until the surface epithelium repairs itself, about 20min. The coverslip piece is removed and the explant transferred to 40 % MSS, where it heals further and then undergoes a characteristic change of shape.

Because these explants are formed from the portion of the marginal zone that has not yet involuted, their composition is very sensitive to the precise stage at which they were made. The lower half of the marginal zone, containing the anterior half of the chordal mesoderm, normally involutes soon after the bottle cells have formed, by stage (Keller et al. 1985).

We chose to create explants at stage instead, for several reasons. Explants can be made much more easily if the yolk mass is well detached from the blastocoel wall. The timing of this event varies from one batch of eggs to the next, but is generally around stage . In addition, we wanted to make dorsal and ventral explants from the same embryos, so it was necessary to wait until the blastopore had reached the ventral side. The implication is that the amount of involuting marginal zone contained in the explant varies from one group to the next (Fig. 1). But convergent extension is a property of both involuting (chordamesoderm) and noninvoluting (neural plate ectoderm) regions, so the main consequence of the absence of the lower involuting marginal zone is that extension begins 1–2 h later than is observed in complete explants (Keller et al. 1985).

Li+ causes progressive axial anteriorization

An empirical scale was developed to facilitate the comparison of the anteriorization of the axis by Li+ (Kao et al. 1986) with its effect on morphogenetic movements. This scale, which distinguishes degrees of loss of posterior structures, is complementary to that used previously to describe the loss of anterior structures in ventralized embryos (e.g. Scharf & Gerhart, 1980). Embryos were scored according to the five grades pictured in Fig. 2:

Fig. 2.

Grades of axial anteriorization, at control stage 42. (A) Grade 0. (B) Grade 1. Note both the oedema and the attenuated tail. (C) Grades 2 (lower), 3 (upper left), and 4 (upper right). Bar, 1 mm.

Fig. 2.

Grades of axial anteriorization, at control stage 42. (A) Grade 0. (B) Grade 1. Note both the oedema and the attenuated tail. (C) Grades 2 (lower), 3 (upper left), and 4 (upper right). Bar, 1 mm.

Grade 0: Normal in all visible respects.

Grade 1: Minor. Some degree of oedema and/or reduction of tail fin size.

Grade 2: Medium. Truncated tail, but somites extend beyond hindgut.

Grade 3: Severe. Normal anterior, but no somites appear beyond hindgut.

Grade 4: Spherical. Head sits atop a spherical ventral cavity. There may be enough somitic tissue for twitching to occur.

Brief treatments in 0·30M-LIC1 dissolved in 40% MSS were chosen over longer treatments in more dilute solutions, because they gave more consistent results. Fig. 3 shows the relationship between the duration of exposure of 64-cell (stage-) embryos to Li+ and the resultant axial anteriorization score. Because intermediate treatment durations produced an extremely variable anteriorization (Fig. 3, inset), no attempt was made to establish the shape of the curve between 3 and 6 min exposure. We did confirm the report (Kao et al. 1986) that later exposure to Li+ causes lesser degrees of axial anteriorization (data not shown).

Fig. 3.

Dose-response relation between axial anteriorization and duration of exposure to Li+. Inset: histogram showing the extreme variability in scores resulting from intermediate exposure duration, in this case 4·5 min. Throughout the text, error bars reflect standard errors of the mean and number of experiments are indicated.

Fig. 3.

Dose-response relation between axial anteriorization and duration of exposure to Li+. Inset: histogram showing the extreme variability in scores resulting from intermediate exposure duration, in this case 4·5 min. Throughout the text, error bars reflect standard errors of the mean and number of experiments are indicated.

Li+ inhibits formation of the blastoporal groove

Blastopore formation begins at the same time in control and Li+-treated embryos. According to morphological criteria, Li+-exposed embryos appear to be divided into the correct subregions necessary to carry out archenteron formation. Bottle cells are present, the leading edge mesoderm is delimited from the overlying marginal zone and this region appears to have the correct thickness (Fig. 4). Nonetheless, formation of the blastoporal groove proper is prevented. While some unusual cell shapes are discernible in Fig. 4, no reproducible alterations were observed.

Fig. 4.

Scanning electron micrograph shows the effect of 6 min Li+ exposure on blastopore formation, at stage 1014. (A) Control. (B) Li+-treated. Note the absence of the blastoporal groove in B. be, bottle cells; iz, involuting marginal zone; Im, leading edge mesoderm. Bar, 50 μm.

Fig. 4.

Scanning electron micrograph shows the effect of 6 min Li+ exposure on blastopore formation, at stage 1014. (A) Control. (B) Li+-treated. Note the absence of the blastoporal groove in B. be, bottle cells; iz, involuting marginal zone; Im, leading edge mesoderm. Bar, 50 μm.

In Li+-treated embryos, the pigment line spreads more rapidly than in controls and its removal from the embryonic surface by blastoporal groove formation is prevented. The result is that the blastopore appears radially symmetrical after 1h at 20°C and retains that appearance throughout gastrulation.

Therefore, to aid in the identification of the dorsal side, the meridian on which the pigment condensation first appeared was marked with Nile blue dye. This mark predicted the direction of archenteron elongation quite well (data not shown).

Li+ causes symmetrization of migration of the leading edge mesoderm

In Xenopus, the mesoderm destined for anteriormost positions is located internally, attached to the base of the endodermal yolk mass, (Nieuwkoop & Florshutz, 1950; Keller, 1976). At stage 10, the dorsal leading edge, consisting of head mesoderm, starts to migrate upward along the noninvoluted mesoderm, lifting the yolky endodermal mass to which it is attached. The movement spreads laterally, reaching the ventral side before stage 11 (Nakatsuji, 1975). We measured leading edge displacement at stage , at which time control embryos exhibit obvious dorsoventral asymmetry, since upward movement of the leading edge is still confined to the dorsal side.

Effects of Li+ upon the dorsoventral pattern of the migration of the leading edge mesoderm are directly reflected in measurements of the dorsal and ventral displacement (Fig. 5). 3 and 6min exposures cause progressive symmetrization of the migration, due primarily to earlier migration of the ventral leading edge. Thus, the cells of the ventral leading edge appear to have been respecihed to migrate as if they were located on the dorsal side. 9min exposure causes no additional symmetrization, but does further delay migration on both sides of the embryo.

Fig. 5.

Dose–response relation between leading edge mesoderm migration and duration of exposure to Li+. Asymmetry is the difference between the dorsal and ventral displacements.

Fig. 5.

Dose–response relation between leading edge mesoderm migration and duration of exposure to Li+. Asymmetry is the difference between the dorsal and ventral displacements.

Li+ retards dorsal involution

We analysed the dorsoventral pattern of involution quantitatively in control and dye-marked Li+-treated embryos when the (late gastrula) controls reached stage 12. This time point occurs after most of the involuting marginal zone has rolled over the blastopore lip, but before gross extension of the marginal zone occurs to alter the overall shape of the embryo. The assessment, therefore, is of the result of the morphogenetic movements which occur during the normal gastrulation period. We did not continue the analysis to later time points, because, in Li+-treated embryos, marginal zone extension generates a number of bizarre embryonic forms (see below), which are difficult to compare geometrically. The process of involution appears to last longer in these embryos, so that by midneurula stage many do exhibit obvious, if unusually shaped, archenterons.

Fig. 6 shows the dramatic retardation of dorsal involution, which is not accompanied by enhancement of movement on the ventral side. Thus, the dorsal involuting marginal zone appears to behave like its ventral counterpart.

Fig. 6.

Dose-response relation between dorsal and ventral involution at stage 12 and duration of exposure to Li+.

Fig. 6.

Dose-response relation between dorsal and ventral involution at stage 12 and duration of exposure to Li+.

We examined the shapes of the archenteron roofs, to see whether Li+ treatment had caused appreciable widening of the archenteron roof. This effect was far less reproducible than the decrease in its length. Formation of radially symmetrical archenterons was observed occasionally.

There was much variability, from one spawning to the next, in the relative degrees of retardation and widening of the shield of involution. Though the general dose-response relationship for retardation of involution correlates well with that for axis effects (Fig. 3), there were a few experiments in which Li+ caused axis effects at doses which did not retard involution much. These observations suggest that conclusions drawn from analysis of archenteron formation would be provisional, at best.

Li+ causes extraordinary tissue movements from late gastrula stage on

After stage 12, extension of the marginal zone causes dramatic lengthening of control embryos. In Entreated embryos, a portion of the involuting marginal zone remaining in the vicinity of the blastopore forms a thick ring of tissue (Fig. 7A) and extends. Using time-lapse video, we have observed that, in cases where the blastopore lips have not met on schedule, they spread rapidly and symmetrically at this time to engulf the yolk plug (data not shown). The dorsal and ventral lips fuse, and extend, generating a proboscis, which may be directed internally or externally (Fig. 7B,C). Extension of the proboscis occurs in lieu of elongation of the body into the characteristic neurula shape.

Fig. 7.

Formation of probosces. (A) Part of the dorsal, d. and ventral, v, marginal zones forms a thick ring around the blastopore, b, at control stage 13. (B) At stage 15, the marginal zone has extended (indicated by the two labelled arrowheads), forming an external proboscis, pressed tightly against the surrounding tissue by the vitelline membrane. Note the blastopore (unlabelled arrowhead) has been squeezed tightly shut. (C) Embryo with an internal proboscis, stage 19. Note the notochord, n and incompletely closed neural tube (unlabelled arrowhead). A and B are midsagittal sections; C is frontal. Bar, 90μm for each.

Fig. 7.

Formation of probosces. (A) Part of the dorsal, d. and ventral, v, marginal zones forms a thick ring around the blastopore, b, at control stage 13. (B) At stage 15, the marginal zone has extended (indicated by the two labelled arrowheads), forming an external proboscis, pressed tightly against the surrounding tissue by the vitelline membrane. Note the blastopore (unlabelled arrowhead) has been squeezed tightly shut. (C) Embryo with an internal proboscis, stage 19. Note the notochord, n and incompletely closed neural tube (unlabelled arrowhead). A and B are midsagittal sections; C is frontal. Bar, 90μm for each.

Internally directed probosces differentiate into notochord and somitic tissue (Fig. 8). Externally directed ones either fall apart or, owing to pressure from the vitelline wall, are reabsorbed into the main embryonic mass.

Fig. 8.

Scanning electron micrograph of fracture through the base of an internal proboscis at late neurula stage. Notochord and somite tissue are distinguishable. n, notochord; s, somitic tissue. Bar, 50 μm.

Fig. 8.

Scanning electron micrograph of fracture through the base of an internal proboscis at late neurula stage. Notochord and somite tissue are distinguishable. n, notochord; s, somitic tissue. Bar, 50 μm.

Sandwich explants show a simple effect of Li+ on convergent extension

In order to simplify interpretation of the effects of Li+ on the process of extension, we made sandwich explants of the marginal zone (Fig. 1). As reported by Keller et al. (1985), control explants from the dorsal side undergo cell rearrangements that correspond well with those in whole embryos and exhibit simultaneous narrowing and lengthening (Fig. 9A,C). Conversely, control ventral explants, like the corresponding region in the embryo, do not narrow or lengthen (Fig. 9B). We studied the behaviour of explants from embryos which had been exposed to Li+ for 6min, a dose ensuring that the majority of individuals would display severe axial anteriorization (Fig. 3), and controls, focusing on the ability of these explants to lengthen.

Fig. 9.

Sandwich explants of the marginal zone. (A) Well-healed dorsal explants, prior to extension. Ventral explants appear identical at this stage. (B) Control ventral explants after extension is complete, 8h later. (C) Control dorsal explants after extension. The one on the left is inverted. (D) Dorsal explants from Li+-treated embryos, after extension. (E) Ventral explants from Entreated embryos, after extension. North–south corresponds to the animal–vegetal axis in the embryo. Bar, 1 mm.

Fig. 9.

Sandwich explants of the marginal zone. (A) Well-healed dorsal explants, prior to extension. Ventral explants appear identical at this stage. (B) Control ventral explants after extension is complete, 8h later. (C) Control dorsal explants after extension. The one on the left is inverted. (D) Dorsal explants from Li+-treated embryos, after extension. (E) Ventral explants from Entreated embryos, after extension. North–south corresponds to the animal–vegetal axis in the embryo. Bar, 1 mm.

Control explants behaved reliably (Table 1). Explants from Li+-soaked embryos exhibited two types of behaviour. 90% (45/50) of the explants extended in a manner typical of controls. 10 % (5/50) of them extended weakly and instead sprouted multiple bizarre stubby ‘antlers’ from the animal half of the explant. The cause of this behaviour is obscure. These explants were excluded from the data.

Table 1.

Effect of Li+ on marginal zone extension

Effect of Li+ on marginal zone extension
Effect of Li+ on marginal zone extension

Fig. 9 and Table 1 show that Li+ confers on ventral explants the ability to extend, behaviour which is never observed in controls. The extension of dorsal explants is not significantly lessened. To ensure that this result could not have arisen from confusion of dorsal and ventral sides of the embryo, we examined the 14 cases in which both dorsal and ventral explants were created from the same pair of embryos. In every one of these, the dorsal explant extended more than the ventral. We conclude that dorsal and ventral embryo halves can be reliably distinguished even in Li+-treated embryos. As a further check, hybrid dorsal-to-ventral explants were created from control embryos. These had a puckered, irregular appearance and were clearly distinguishable after extension by a crenelated surface which contrasted dramatically with the smooth surface of all other types of explants.

Single batches of eggs in which the yolk mass separated especially easily from the blastocoel wall yielded explants displaying strikingly homogeneous behaviour. Fig. 10 shows the results of an exemplary experiment, in which the ventral explants extend to 60% of the dorsal value. Also, it appears that explants from Li+-soaked embryos extend on the same schedule as those from controls.

Fig. 10.

Time course of explant extension, in a representative experiment. (■) Control dorsal (5 explants). (▴) Control ventral (4). (•) Li+-treated dorsal (5). (▼) Li+-treated ventral (4).

Fig. 10.

Time course of explant extension, in a representative experiment. (■) Control dorsal (5 explants). (▴) Control ventral (4). (•) Li+-treated dorsal (5). (▼) Li+-treated ventral (4).

Li+ as a dorsalizing agent

We have described three major effects of Li+ on the morphogenetic movements of gastrulation: (1) symmetrization of the migration of the leading edge mesoderm at the onset of gastrulation, due to precocious movement of the ventral region, (2) retardation of involution of the chordal mesoderm and the associated archenteron roof and (3) conferral of convergent extension upon the ventral marginal zone. We can interpret only two of these effects as dorsalization.

Sandwich explants clearly show that ventral positional information has been respecified to more dorsal values. The ventral marginal zone is caused to undergo convergent extension (Keller et al. 1985), which normally occurs only on the dorsal side, while the behaviour of the dorsal zone is unaffected (Table 1). Because Li+ causes whole embryos to exhibit such a variety of form, we were quite gratified to find that sandwich explants show a dorsalizing effect of Li+ in such a simple way.

In the case of movement of the leading edge mesoderm, Li+ clearly causes precocious ventral migration (Fig. 5). There is no significant retardation of dorsal migration at 6 min Li+ exposure (P ≃0-5 using an unpaired t-test), a dose adequate to ensure near maximal anteriorization (Fig. 3) and pronounced ventral extension (Table 1). We also note that increasing Li+ exposure from 6–9 min causes parallel retardation of both dorsal and ventral leading edge movement, so that the asymmetry is unchanged (Fig. 5). We conclude, therefore, that this retardation is a distinct, high-dose effect of Li+. At lower doses, leading edge migration shows respecification of ventral mesoderm to more dorsal identity.

We have made a number of separate observations which bear on the interpretation of the effect of Li+ on dorsal involution (Fig. 6). In a single experiment, 6 min exposure to Li+ caused pronounced axial anteriorization in embryos which exhibited normal archenteron formation. In general, increasing Li+ exposure from 6 to 9 min causes pronounced further inhibition of involution, despite the fact that near maximal axial anteriorization occurs after a 6min exposure. These two clues suggest that the correlation between retardation of involution and axial anteriorization might be coincidental. Also, 9 min Li+ exposure often, and lower doses sometimes, caused some embryos to be unable to accomplish any involution whatsoever. They failed to enclose the yolk mass and soon lysed. These embryos probably represent extreme examples of mechanical disorganization at the blastopore which, in milder cases, only delays the process of involution.

We often find very long bottle cells in Li+-treated embryos. Because explanted bottle cells do not retain their elongate shape, Keller (1981) concluded that this shape reflects the external force applied to the cells. Thus, overlong bottle cells might reflect excess tension in the region. If upward migration of the leading edge mesoderm were to pull the bottle cells, while the edge of the involuting marginal zone, to which their apical surfaces are firmly attached, was prevented from turning in, the bottle cells would be stretched.

It would be desirable to use sandwich explants to learn more about the reason involution is prevented. However, we have not been able to create explants from Li+-treated embryos which reliably include the lower marginal zone. This suggests that there is very real mechanical disorganization at the blastopore, despite the lack of clear signs of this in electron micrographs.

Despite the retardation of involution, we did observe some degree of angular broadening of the archenteron roof and thus of respecification of ventral mesoderm to more dorsal values. In the absence of additional observations, the characteristic truncated appearance of the roof might suggest that both dorsal and ventral positions were behaving as intermediate, lateral ones, i.e. that the positional gradient was being flattened. However, the whole host of observations about the effect of Li+ on involution suggest that truncation of the archenteron roof arises from a mechanical disruption at the blastopore which is distinct from the alteration of dorsoventral positional values.

Nieuwkoop (1970) has shown that Li+ enhances the receptivity of blastula animal cap cells to mesodermal inductive signals from the yolky endoderm. Retardation of involution might then arise from overinduction of the marginal zone. A signal propagating upward from the yolky endoderm would be strongest at the lower extreme of the marginal zone, which is the site of the blastopore, and gradually decay through the involuting marginal zone, noninvoluting marginal zone and animal cap regions. Assuming that bottle cells arise from a maximal induction, we would predict that excess bottle cells should form. Unfortunately, we have not observed this phenomenon to date.

In normal gastrulation, movement of the involuting marginal zone is continuous with the movement of the internal leading edge mesoderm which precedes it (Keller, 1976). Yet the motile behaviour of these cell groups is quite different; the leading edge cells migrate as individuals or small groups (e.g. Kubota & Durston, 1978), while the involuting marginal zone moves as a cell sheet (Keller, 1986). Li+-induced disorganization of the blastopore disrupts involution without inhibiting leading edge migration and thus functionally distinguishes the two cell groups.

Proposals for an assay

The normal course of morphogenesis seems to require more-precise spatial organization than does subsequent axial differentiation. In contrast to our expectations, the shape of the archenteron proved to be a poor indicator, indeed, of the region-specific pattern of cell behaviour. This implies that extreme caution must be used in attempting to infer cellular properties from morphogenetic movements of tissues. There is some precedent for this view. Scharf & Gerhart (1980) noted that though gravity-driven cytoplasmic rearrangement could rescue axis formation in u.v.-irradiated embryos, blastopore formation and closure remained like that in the ventralized ones.

This work was undertaken in order to develop a quantitative assay for alterations to the dorsoventral pattern. Measures based on scoring of form, like the axial anteriorization scale, are most useful when large groups of embryos can be compared. If pattern formation is to be investigated using manipulations performed on small numbers of embryos, such as microinjection or blastomere transplantation, an assay that yields statistically significant results from small groups of specimens is desirable. As we have shown, the pattern of morphogenesis in whole embryos can be extremely misleading. On the other hand, sandwich explants of the marginal zone appear to be ideal for this purpose. 20 experimental embryos should easily suffice to make five good ventral explants. The length to which these explants extend is a quantitative and direct measure of the degree to which ventral positional values have been dorsalized.

We are grateful to Ray Keller for many invaluable discussions, for use of laboratory facilities, and particularly for showing one of us (C.M.R.) how to make sandwich explants of the marginal zone. This work was supported by NSF grant PCM 84-020892 to R.A.S. C.M.R. is an NIH Systems and Integrative Biology Trainee.

Bäckström
,
S.
(
1954
).
Morphogenetic effects of lithium on the embryonic development of Xenopus
.
Arkiv. Zool
.
6
,
527
536
.
Gerhart
,
J. C.
&
Keller
,
R.
(
1986
).
Region-specific cell activities in amphibian gastrulation
.
A. Rev. Cell Biol
.
2
,
201
229
.
Gerhart
,
J. C.
,
Vincent
,
J.-P.
,
Scharf
,
S. R.
,
Black
,
S. D.
,
Gimlich
,
R. L.
&
Danilchik
,
M.
(
1984
).
Localization and induction in early development of Xenopus
.
Phil. Trans R. Soc. Land. B
307
,
319
330
.
Gimlich
,
R. L.
(
1986
).
Acquisition of developmental autonomy in the equatorial region of the Xenopus embryo
.
Devi Biol
.
115
,
340
352
.
Gimlich
,
R. L.
&
Gerhart
,
J. C.
(
1984
).
Early cellular interactions promote embryonic axis formation in Xenopus laevis
.
Devi Biol
.
104
,
117
130
.
Grant
,
P.
&
Wacaster
,
J. F.
(
1972
).
The amphibian grey crescent - a site of developmental information?
Devi Biol
.
28
,
454
468
.
Gurdon
,
J. B.
(
1987
).
Embryonic induction-molecular prospects
.
Development
99
,
285
306
.
Humason
,
G. L.
(
1979
).
Animal Tissue Techniques
.
San Francisco
:
W. H. Freeman and Co
.
Kao
,
K. R.
,
Masui
,
Y.
&
Elinson
,
R. P.
(
1986
).
Lithium-induced respecification of pattern in Xenopus laevis embryos
.
Nature, Land
.
322
,
371
373
.
Keller
,
R. E.
(
1976
).
Vital dye mapping of the gastrula and neurula of Xenopus laevis. II. Prospective areas and morphogenetic movements of the deep layer
.
Devi Biol
.
51
,
118
137
.
Keller
,
R. E.
(
1981
).
An experimental analysis of the role of bottle cells and the deep marginal zone in gastrulation of Xenopus laevis
.
J. exp. Zool
.
216
,
81
101
.
Keller
,
R. E.
(
1986
).
The cellular basis of amphibian gastrulation
.
In Developmental Biology, A Comprehensive Synthesis
(ed.
L. W.
Browder
), pp.
241
327
.
New York
:
Plenum Press
.
Keller
,
R. E.
,
Danilchik
,
M.
,
Gimlich
,
R.
&
Shih
,
J.
(
1985
).
The function and mechanism of convergent extension during gastrulation of Xenopus laevis
.
J. Embryol. exp. Morph
.
89
Supplement
,
185
209
.
Kirschner
,
M.
&
Hara
,
K.
(
1980
).
A new method for local vital staining of amphibian embryos using Ficoll and “Crystals” of Nile Red
.
Mikroskopie
36
,
12
15
.
Kubota
,
H.
&
Durston
,
A. J.
(
1978
).
Cinematographical study of cell migration in the opened gastrula of Ambystoma mexicanum
.
J. Embryol. exp. Morph
.
44
,
71
80
.
Malacinski
,
G. M.
,
Brothers
,
A. J.
&
Chung
,
H.-M.
(
1977
).
Destruction of components of the neural induction system of the amphibian egg with ultraviolet irradiation
.
Devi Biol
.
56
,
24
39
.
Nakatsuji
,
N.
(
1975
).
Studies on the gastrulation of amphibian embryos: cell movement during gastrulation in Xenopus laevis embryos
.
Wilhelm Roux’ Arch, devl Biol
.
178
,
1
14
.
Nieuwkoop
,
P. D.
(
1970
).
The formation of the mesoderm in urodelean amphibians. III. The vegetalizing action of the Li+ ion
.
Wilhelm Roux’ Arch, devl Biol
.
166
,
105
123
.
Nieuwkoop
,
P. D.
(
1973
).
The “organization center” of the amphibian embryo: its origin, spatial organization, and morphogenetic action
.
Adv. Morph
.
10
,
1
39
.
Nieuwkoop
,
P. D.
&
Faber
,
J.
(
1967
).
Normal Table of Xenopus laevis (Daudin)
.
Amsterdam
:
North-Holland
Nieuwkoop
,
P.
&
Florshutz
,
P.
(
1950
).
Quelques caractères spéciaux de la gastrulation et de la neurulation de l’ouef de Xenopus laevis, Daud. et de quelques autres Anoures, l’ère partie. Étude déscriptive
.
Archs Biol. Liège
61
,
113
150
.
Regen
,
C. M.
&
Steinhardt
,
R. A.
(
1986
).
Global properties of the Xenopus blastula are mediated by a high-resistance epithelial seal
.
Devl Biol
.
113
,
147
154
.
Scharf
,
S. R.
&
Gerhart
,
J. C.
(
1980
).
Determination of the dorsal-ventral axis in eggs of Xenopus laevis: complete rescue of uv-impaired eggs by oblique orientation before first cleavage
.
Devl Biol
.
79
,
181
198
.
Scharf
,
S. R.
&
Gerhart
,
J. C.
(
1983
).
Axis determination in eggs of Xenopus laevis: a critical period before first cleavage, identified by the common effects of cold, pressure and ultraviolet irradiation
.
Devl Biol
.
99
,
75
87
.
Scharf
,
S. R.
,
Vincent
,
J.-P.
&
Gerhart
,
J. C.
(
1984
).
Axis determination in the Xenopus egg
.
UCLA Symp. molec. cell. Biol
.
19
,
51
73
.
Slack
,
J. M. W.
(
1983
).
From Egg to Embryo: Determinative Events in Early Development
.
Cambridge, London
:
Cambridge University Press
.
Spemann
,
H.
(
1938
).
Embryonic Development and Induction
.
Yale Univ. Press
. Reprinted 1967.
New York
:
Hafner Publishing Co
.
Suzuki
,
A. S.
,
Mifune
,
Y.
&
Kanéda
,
T.
(
1984
).
Germ layer interactions in pattern formation of amphibian mesoderm during primary embryonic induction
.
Dev. Growth Differ
.
26
,
81
94
.
Vincent
,
J.-P.
,
Oster
,
G. F.
&
Gerhart
,
J. C.
(
1986
).
Kinematics of gray crescent formation in Xenopus eggs: the displacement of subcortical cytoplasm relative to the egg surface
.
Devl Biol
.
113
,
484
500
.